https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127689
--- Comment #3 from Drea Pinski <pinskia at gcc dot gnu.org> --- Created attachment 65740 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=65740&action=edit Semi reduced Yes it looks like I was correct. I have to check but I think there is already a pattern which handles the truncation for integer types. So this pattern should be only for FP types. and also TYPE_PRECISION should be changed to element_precision for vector types (conversions don't show up much for vectors which is why the ICE has not happened yet).
